This application documents several lines of evidence which suggest that cell-cell interactions play a significant role in erythropoiesis and that these interactions are amenable to investigation in vitro employing proliferation dependent assays for erythroid colony forming cells in plasma clots. Thus, we propose to obtain discrete populations of mouse bone marrow cells by utilizing various fractionation procedures, or combinations of procedures, including unit gravity sedimentation, electrophoresis, bouyant density centrifugation, adherence techniques, and immune lysis of particular cells. The populations of marrow cells obtained by these methods will be cultured alone, in combination with other cell fractions, or with unfractionated bone marrow cells, and the number of bursts of erythroid colonies (BFU-E) and of erythroid colonies (CFU-E) formed in response to appropriate concentrations of erythropoietin determined. A synergistic, rather than additive, increase in the number of colonies produced will be indicative of erythropoietically significant cell-cell interaction. Although several of these fractionation methods are known to result in populations of cells which influence the development of erythroid colonies in culture, an additional approach to this problem consists of employing agents, both in vivo and in vitro, which selectively enhance or suppress the activity of particular cells (e.g., macrophages and/or lymphocytes); in other experiments, we will employ mice which are genetically defective in certain lymphocyte populations (e.g., nude mice). Erythroid colony formation by bone marrow cells obtained from these mice or cultured in the presence of the agents may also serve to demonstrate, inferentially, the role of cell-cell interaction in erythropoiesis.